import heapq

def dijkstra2(adjList, s, g):
    # Initialize the distance from the start node s to all other nodes as infinity
    distance = [float('inf')] * len(adjList)
    distance[s] = 0

    # Initialize the priority queue with the start node
    queue = [(0, s)]  # (distance, node)

    while queue:
        # Get the node with the smallest distance
        dist, node = heapq.heappop(queue)

        # If this node has already been processed, then ignore it
        if dist > distance[node]:
            continue

        # Update the distances to the neighboring nodes
        for neighbor, cost in adjList[node]:
            new_dist = dist + cost
            if new_dist < distance[neighbor]:
                distance[neighbor] = new_dist
                heapq.heappush(queue, (new_dist, neighbor))

    # If the distance to the goal node g is infinity, it means there is no path.
    if distance[g] == float('inf'):
        return 1e10
    else:
        return distance[g]


N, M = map(int, input().split())

adj_list = [[] for _ in range(N)]
for i in range(M):
    u, v = map(int, input().split())
    u -= 1
    v -= 1
    adj_list[u].append((v, 1))

ans1 = dijkstra2(adj_list, 0, N-2) + dijkstra2(adj_list, N-2, N-1) + dijkstra2(adj_list, N-1, 0)
ans2 = dijkstra2(adj_list, 0, N-1) + dijkstra2(adj_list, N-1, N-2) + dijkstra2(adj_list, N-2, 0)

ans = min(ans1, ans2)
if (ans >= 1e10):
    print(-1)
else:
    print(ans)